Isolation and complete amino acid sequence of osteocalcin from canine bone

Osteocalcin was purified in high yield and to homogeneity from the diaphysis of dog femora by the following steps: (1) acid demineralization of bone powder, (2) solid‐phase extraction of acid‐soluble proteins on Sep‐Pak C₁₈ cartridges, (3) gel filtration on Sephadex G‐50, and (4) fast protein liquid chromatography on an Accell‐QMA anion‐exchange column. Starting from 30 g washed bone powder, approximately 7–10 mg pure protein was obtained in 2 days. The key step is the initial solid‐phase extraction of osteocalcin from a large volume of a demineralized bone solution. The primary structure was established by automated sequence analyses of two tryptic peptides, of two endoproteinase Glu‐C carboxy‐termina) peptides, and of the first 30 amino acid residues of the intact protein. Dog osteocalcin contains 49 amino acids, has a molecular mass of 5654 daltons, contains no Thr, Met, Hyp, or Trp, has a disulfide bond between Cys 23 and 29, and is fully γ‐carboxylated at residues 17, 21, and 24. Dog osteocalcin does not contain a pair of basic amino acids found at positions 43–44 in most other osteocalcins from mammals and birds. A computer search for homology indicated 88, 90, 84, 88, 66, and 57% sequence identity of dog osteocalcin with human, bovine, cat, monkey, chicken, and swordfish osteocalcin, respectively, and weaker homologies with the γ‐carboxylated domains of blood‐clotting proteins and the Pro‐rich N‐terminal extensions of myosin light‐chain A1 and β‐crystalline B1. The possible relevance of these homologies to the structure and potential functions of osteocalcin is discussed.